The fusion framework in a person identity verification system based on face and speech data

A personal verification task performed by an expert system is a binary classification problem which consists of accepting or rejecting an identity claim. An automatic system for the person identity authentication based only on one type of data is often unable to meet the required performance. The combination of multiple biometric attributes such as, face data, speech, should result in a more reliable and accurate system. By comparing the recent methods used in the identity verification, a framework for the next research challenges is provided.